metadata
dataset_info:
features:
- name: arxiv_id
dtype: string
- name: published
dtype: string
- name: titles
dtype: string
- name: authors
sequence: string
- name: abstract
dtype: string
- name: categories
sequence: string
- name: selected
dtype: bool
splits:
- name: train
num_bytes: 52422214
num_examples: 35441
- name: valid
num_bytes: 7883002
num_examples: 5273
- name: test
num_bytes: 7661672
num_examples: 5105
download_size: 39071671
dataset_size: 67966888
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*
- split: test
path: data/test-*